Does Percentage of Latinas/os Affect Graduation Rates at 4-Year Hispanic Serving Institutions (HSIs),Emerging HSIs,and Non-HSIs?

Abstract

As the population of college-aged Latinas/os grows, the number of Hispanic Serving Institutions (HSIs) increases. The purpose of this study was to determine whether the percentage of Latinas/os has an effect on the institutional graduation rates of Latina/o students attending HSIs, emerging HSIs, and non-HSIs. Data were drawn from the Integrated Postsecondary Education Data System (IPEDS) and the sample included 296 institutions. A structural equation model was used to confirm predictors of graduation rates.

Keywords

Hispanic Serving Institutions (HSIs)emerging HSIs Latina/o college students graduation rates

As the population of Latinas/os increases, the number attending college is also increasing. Similarly, the influx of Latina/o college students is ultimately leading to a growth in the number of Hispanic Serving Institutions (HSIs). Defined by the 1998 reauthorization of the Higher Education Act (HEA), HSIs are accredited, degree-granting institutions that have at least 25% full-time equivalent (FTE) enrollment of undergraduate Latina/o students (Laden, 2004). Currently, HSIs represent approximately 8% of all colleges and universities but they enroll more than 50% of all Latina/o college students (White House Initiative on Educational Excellence for Hispanics, 2011). In addition, Santiago and Andrade (2010) report that in 2006-2007, there were 176 “emerging” HSIs with a critical mass of 15-24% Latina/o student enrollment. Research on HSIs and emerging HSIs, however, is scant. The purpose of this study was to determine if the percentage of Latina/o students, faculty, and staff at HSIs, emerging HSIs, and non-HSIs has an effect on graduation rates.

Literature Review

Persistence and Graduation at HSIs

HSIs are responsible for conferring a large portion of the degrees awarded to Latina/o students in the United States (Mercer & Stedman, 2008). In fall 2001-2002, HSIs awarded 55.6% of all associates degrees to Latina/o students and 42% of all bachelor’s degree (Mercer & Stedman, 2008). Although these statistics suggest that HSIs may be effectively graduating Latina/o students, a number of scholars have looked beyond the raw numbers to determine which factors have an effect on the success of Latina/o students attending HSIs.

Hagedorn, Chi, Cepeda, and McLain (2007) investigated the importance of Latina/o critical mass on the success of Latina/o community college students. Rather than looking at graduation rates, transfer rates, or degrees conferred, they conceptualized success to include multiple measures including course completion ratios, cumulative GPA, and math and English course completion. They found that attitude, aspirations, and representational value of Latina/o students increased the likelihood of Latina/o students’ success (Hagedorn et al., 2007). When they substituted students representational value for faculty representational value, they also found a statistically significant effect on success (Hagedorn et al., 2007). Despite the positive association between critical mass and success, the findings in their study are relative to the way success was operationalized.

In reconceptualizing a model of persistence for Latina/o college students, Torres (2006) used a combination of interviews and survey data collected from Latina/o students at three urban institutions, two of which were HSIs. Using a structural equation model, she found that Cultural Affinity had a direct effect on academic integration and encouragement, which had a direct effect on institutional commitment and an indirect effect on intent to persist (Torres, 2006). Cultural Affinity was conceptualized as a latent construct consisting of students’ responses to three questions: (a) Latino faculty and staff help me to feel at home at this college, (b) Other Latino students help me to feel at home at this college, and (c) Latino cultural activities help me to feel at home at this college (Torres, 2006). Cultural Affinity could be unique to HSI environments where Latino faculty and staff are likely to be represented in larger numbers and could ultimately enhance persistence at these institutions.

To determine the value added context for Latinas/os attending HSIs, Contreras, Malcolm, and Bensimon (2008) used data from the Fall 2004 Integrated Postsecondary Education Data System (IPEDS) to explore degree production at five 4-year HSIs. They found that all five 4-year HSIs in their sample were below equity in degree production for Latinas/os compared with White students who were above equity in four of the five institutions (Contreras et al., 2008). Although these findings are limited due to a sample size of five purposefully selected HSIs, the equity concern that Contreras et al. (2008) raise is whether the number of degrees conferred to Latinas/os is comparable to the total number of degrees conferred to all students, which is an important consideration.

Malcom (2010) assessed the institutional performance of 4-year HSIs (enrolling 25% or more Latina/o students), emerging HSIs (enrolling 15%-24% Latina/o students), and non-HSIs (enrolling less than 15% Latina/o students). Using a sample of 143 4-year institutions and IPEDS data from 2006-2007 and 2007-2008, she concluded that 4-year HSIs with Latina/o enrollment above 33% had lower 6-year graduation rates than non-HSIs and emerging HSIs. These same institutions, however, outperformed non-HSIs and emerging HSIs in a number of other areas including employing a larger proportion of Latina/o faculty and administrators (Malcom, 2010). The mixed results of these studies suggest that further exploration is needed of institutional graduation rates for Latinas/os attending HSIs, emerging HSIs, and non-HSIs.

Institutional Graduation Rates

Graduation rates are often used as indicators of postsecondary student success and may be one indicator of effectiveness for institutions. Several variables have been determined to be important indicators of institutional graduation rates and should be considered when analyzing these rates. Selectivity, often measured by average freshmen SAT scores and number of students admitted (Alon & Tienda, 2005; Bowen & Bok, 1998), has emerged as one major institutional factor that affects the graduation and success of students of color. Bowen and Bok (1998), for example, found that for both Black and White students with similar SAT scores, attending a more selective institution increases graduation rates. Several researchers (Alon & Tienda, 2005; Kane, 1998; Melguizo, 2008) found similar results, thus concluding that institutional selectivity increases the likelihood that students of color will graduate.

Researchers have also analyzed the effects of institutional funding and expenditures on institutional graduation rates. Ryan (2004) used IPEDS data to examine the effect of four expenditure variables on institutional graduation rates and concluded that after controlling for student-level inputs, instructional expenditures and academic support expenditures significantly contributed to the variance in 6-year graduation rates of students at baccalaureate colleges. Gansemer-Topf and Schuh (2006) found that in addition to expenditures on instruction and academic support, administrative support expenditures per student and institutional grant expenditures per student significantly affected 6-year graduation rates. Zhang (2009) concluded that for public institutions across Carnegie Classification types, an increase in state appropriations per FTE was related to an increase in institutional graduation rates. Shin (2010) also found that institutional-level variables, including faculty-student ratio, instructional expenditures per student, campus living facilities, and in-state tuition, accounted for a majority (76%) of the variance in institutional graduation rates. These findings highlight the importance of including institutional selectivity and various institutional expenditures in an exploration of institutional graduation rates.

One idea that has not been extensively researched is the relationship between the percentages of Latina/o students and personnel and institutional graduation rates for Latina/o students. As suggested by Malcom (2010), many HSIs outperform non-HSIs in regard to serving Latina/o students, which she determined was partially due to the higher percentage of faculty and administrators that identify as Latina/o at HSIs. Hurtado and Kamimura (2003) contend that increasing the number of Latina/o faculty and administrators on campus is essential for increasing the retention of Latina/o students in postsecondary education because they provide role modeling and mentorship and may help to alleviate feelings of alienation and marginalization. This variable, however, should be explored further, especially for HSIs that are enrollment driven institutions.

Method

The purpose of this study was to determine whether the percentages of Latina/o students, staff, and faculty are related to the graduation rates of Latina/o students attending HSIs, emerging HSIs, and non-HSIs. Structural equation modeling (SEM) was the preferred method because it takes a confirmatory approach to testing theories and analyzing data through nonexperimental procedures (Byrne, 2008). Unlike more exploratory methods, SEM allows the researcher to specifically test whether a hypothesized model of relationships fits the data (Mulaik, 2009).

Data Source

Cross-sectional data from IPEDS were used for this study because IPEDS is a universe sample of all higher education institutions that receive federal financial assistance under Title IV of the HEA of 1965. As a universe sample, data are included for a large range of institutions including public, private, not-for-profit, vocational, 2-year, and 4-year. The data provided by IPEDS are aggregated and can be used to describe and analyze trends in higher education. For this study, data were drawn from the human resources, finance, and graduation rates surveys for fall 2002 and spring 2008.

Sample

The sample of 296 4-year institutions includes HSIs, emerging HSIs, and non-HSIs. In this study, HSIs are defined strictly by the FTE population of Latina/o students, regardless of official designation as a HSI by the Department of Education. The sample of institutions identified as HSIs (n = 62), therefore, includes the entire population of 4-year institutions enrolling 25% or more Latina/o students in fall 2002. The institutions in Puerto Rico were excluded because they enroll nearly 100% Latina/o students and are arguably different than U.S. mainland HSIs because of the cultural and linguistic differences (Contreras et al., 2008). The 62 institutions are heterogeneous by size, control, Carnegie Classification, and percentage of Latina/o enrollment.

The 4-year institutions identified as emerging HSIs (n = 104) in this study enrolled between 10% to 24% Latina/o students in fall 2002. Although previous authors have identified emerging HSIs as those that enroll between 15% to 24% Latina/o students (Malcom, 2010; Santiago & Andrade, 2010), the emerging HSI category is not an official designation and therefore the percentage is arbitrary. For this study, I used 10% as a cutoff point because I used fall 2002 data to define the population and spring 2008 data for the dependent variable. By spring 2008, a majority of the institutions that enrolled between 10% to 14% Latina/o students in fall 2002 had in fact reached the 15% mark. Like the HSIs, the emerging HSIs in this study are heterogeneous by size, control, Carnegie Classification, and percentage of Latina/o enrollment. The HSIs and emerging HSIs were matched with similar 4-year non-HSIs to provide a comparable sample of institutions. The non-HSIs (n = 130) enrolled less than 10% Latina/o students in fall 2002.

Variables

The dependent variable was a single item observed variable that indicates the 6-year institutional graduation rate for Latina/o undergraduate students in 2008. Three single item observed variables were also included as independent variables in the model. The first, standardized test scores, was used as a proxy for institutional selectivity. The SAT math 75th percentile and SAT critical reading 75th percentile aggregated scores were combined to determine one SAT composite 75th percentile aggregated score. For schools that only reported ACT scores, the ACT composite 75th percentile was converted to the equivalent SAT composite 75th percentile aggregated score. The second single item variable included in the model was percentage of Latina/o students enrolled. This item was added to the model to help determine the extent to which HSI status predicts institutional graduation rates for Latina/o students. The final single item variable, institutional control, was included as a control variable in the model because it has previously been found to influence graduation rates (Melguizo, 2008; Ryan, 2004). Institutional control was dummy coded with the referent group “private.”

Two latent variables were also included in the model. The first, institutional resources, was comprised of four observed variables with high factor loadings including instructional expenditures, academic support expenditures, student services expenditures, and institutional support expenditures. These four types of expenditures, as reported in dollar amounts, were chosen because previous studies have indicated that they are good predictors of graduation (Gansemer-Topf & Schuh, 2006; Shin, 2010; Zhang, 2009). The total dollar amount in each category was divided by FTE; therefore, the model includes expenditures per student as opposed to total dollar amounts. The factor loadings for the individual items ranged from .714 to .951 and the institutional resources scale had a Cronbach’s alpha of .819.

The second latent variable was a measure of representation of Latina/o personnel as indicated by the percentage of Latina/o faculty, administrators, and professionals employed at each institution. As suggested by Malcom (2010), many HSIs employ a large number of Latina/o faculty and staff, which increases their effectiveness in serving Latina/o students. For this study, the latent variable was comprised of three observed variables including percentage of Latina/o faculty, percentage of Latina/o administrators, and percentage of Latina/o professionals. The factor loadings for the individual items ranged from .894 to .935 and the full scale had a Cronbach’s alpha of .930. The means, standard deviations, and reliability estimates are presented in Table 1.

Table 1.

Means, Standard Deviations, and Reliability Estimates for Observed Variables.

Variable	M	SD	Alpha
Latina/o 6-year graduation rate	0.45	0.22
Standardized test scores	11.38	1.06
Percentage of Latina/o students enrolled	0.17	0.18
Institutional control	1.55	0.50
Latina/o representation
Percentage of Latina/o faculty	0.06	0.11	.895
Percentage of Latina/o administrators	0.09	0.15	.933
Percentage of Latina/o professionals	0.11	0.16	.931
Institutional resources
Instructional expenditures	0.95	1.21	.929
Academic support expenditures	2.58	4.14	.951
Student services expenditures	2.56	2.04	.720
Institutional support expenditures	4.40	6.36	.926

Data Analysis

As a first step in the analysis, I used SPSS 18 to conduct a missing values analysis (MVA), which revealed that a majority of the missing values were from the standardized test scores variable, with 18.1% of the cases missing this value. This was not surprising because many of the HSIs in the sample are broad access institutions that do not require standardized test scores for admission. To preserve the number of HSIs in the sample, I used the Expectation Maximization (EM) algorithm to replace missing data because it is a more robust technique than mean replacement (Allison, 2002; Dempster, Laird, & Rubin, 1977; McLachlan & Krishnan, 1997).

Using EQS 6.2 for Windows, SEM was then used to determine which variables influence institutional graduation rates for Latina/o students. SEM was an appropriate method because it allows the researcher to simultaneously analyze observed and unobserved variables while correcting for error variance in the model parameters (Byrne, 2008). EQS 6.2 provides a number of model fit indices including the χ² significance test, the comparative fit index (CFI), and the root mean-square error of approximation (RMSEA). As suggested by Hu and Bentler (1999), two recommended values include a CFI value greater than .95 and a RMSEA value below .06.

Prior to checking the fit of the hypothesized model, the data were tested for normality. Mardia’s normalized coefficient = 191.95 was high, indicating nonnormality as a result of multivariate kurtosis. Examining the univariate statistics revealed high kurtosis for several of the observed variables. Rather than eliminating individual cases that were contributing to kurtosis, I used the Maximum Likelihood (ML), robust method offered in EQS 6.2, which obtains ML estimates and adjusts standard errors and test statistics for nonnormality found in the data (Bentler, 2006). All the reported statistics, therefore, are based on the robust method.

Limitations

Although this study examined institutional factors that influence institutional graduation rates, an institution’s graduation rates are also influenced by a number of student characteristics. The use of IPEDS data, however, limits the number of student variables that can be used in the model because few student-level variables are provided. In addition, the student-level data are aggregated, which could be problematic and may increase the chance of committing a Type I error. This study, therefore, is limited in its ability to explain variation in 6-year graduation rates for Latina/o students. Furthermore, a major limitation in using IPEDS institutional graduation rates is that they only account for graduation of first-time, full-time freshmen; therefore, the true graduation rate for Latina/o students may be underestimated as many transfer or follow nontraditional paths through college. This study is also limited because it does not account for precollege institutional predictors of graduation such as high school location, ratings, and racial segregation.

The use of SEM also has limitations to consider. First, the model is limited to providing information based on the researcher’s design. Careful thought and consideration must be given to account for a number of extraneous factors that may be contributing to the variance in the outcome variable of concern. For this study, there are numerous factors that may contribute to graduation rates, both at the student level and at the institutional level, but the design was limited by the data source. In addition, the sample size constricts the extent of the analysis performed in SEM because a small sample size may be limited in power (MacCallum, Browne, & Sugawara, 1996). This limitation made it difficult to study HSIs and emerging HSIs as individual groups. Finally, although Bentler (1988) argues that SEM takes a confirmatory approach to analyze a theory that represents a causal process, the findings ultimately represent the relationship between variables.

Results

The final structural equation model yielded a Satorra-Bentler χ² = 33.42, df = 29, p = .26; CFI = .97; RMSEA = .02, all of which suggest a good-fitting model for the proposed data. The accepted model explained 34% of the variance in graduation rates for Latina/o students with statistically significant paths from standardized test scores (p < .05), institutional resources (p < .05), and institutional control (p < .05). The representation of Latina/o personnel and the percentage of Latina/o students, however, did not have a significant influence on the graduation rates for Latina/o students. The accepted model is displayed in Figure 1.

Figure 1.

Accepted model: Institutional graduation rates for Latina/o students.

Discussion

The purpose of this study was to determine the extent to which percentage of Latina/o personnel and percentage of Latina/o students enrolled predict institutional graduation rates for Latina/o students at HSIs, emerging HSIs, and non-HSIs while controlling for important institutional variables that have been found to be predictive of graduation rates. Previous studies have determined that selectivity and institutional resources are significant predictors of graduation rates for all college students. SEM allowed me to confirm the importance of these variables in regard to predicting Latina/o graduation rates while exploring two variables that are unique to HSIs and emerging HSIs, mainly, the higher population of Latina/o students and personnel on campus.

The final model confirmed the significance of institutional selectivity on graduation rates for Latina/o students. The direct effect of institutional selectivity on Latina/o graduation rates is not surprising, considering the abundance of research that claims that more selective institutions have higher graduation rates for students of color (Alon & Tienda, 2005; Bowen & Bok, 1998; Cragg, 2009; Melguizo, 2008). The importance of selectivity, however, highlights the challenge that educators face in increasing the graduation rates of Latina/o students because they continue to score lower on standardized tests than their White and Asian counterparts (Gandara & Lopez, 1998). If Latina/o students enter college less prepared and less competitive than White and Asian students, additional services must be provided to these students to ensure their academic success and graduation. Lower institutional selectivity cannot be an excuse for lower 6-year graduation rates for Latina/o students.

The final model also confirmed the significance of institutional resources on influencing institutional graduation rates for Latina/o students. To increase graduation rates for Latina/o students, therefore, instructional expenditures, academic support expenditures, student support expenditures, and institutional support expenditures, specifically, must be maintained and/or increased. This is an important finding because de los Santos and Cuamea (2010) argue that presidents of HSIs list declining institutional resources as a main challenge facing their institutions. If key educational resources continue to decrease, HSIs will struggle to retain and graduate the increasing number of Latina/o students entering their institutions. The findings of this study suggest that to increase graduation rates for Latina/o students, it will become more important for institutions of higher education to seek out financial resources that will support this growing population.

The representation of Latina/o personnel was not significantly related to the institutional graduation rates for Latina/o students in this study. In addition, the percentage of Latina/o students enrolled on campus was insignificant in predicting graduation rates. These two variables, however, were of utmost importance to this study because HSIs, by nature, have a larger percentage of Latina/o students and ultimately have a larger percentage of Latina/o personnel on campus. The fact that the percentage of Latina/o students enrolled is not significantly related to graduation rates for these students may be due to the fact that increasing the number of diverse students on campus is not enough to increase educational outcomes for these students (Hurtado, Milem, Clayton-Pedersen, & Allen, 1998, 1999). As argued by Hurtado et al. (1998, 1999) the historical legacy of inclusion must also be considered as well as the psychological and behavioral dimensions of the campus. Institutions, therefore, must think beyond the structural diversity of their campus and start to analyze other aspects of the campus climate that may be related to the graduation of Latina/o students. Simply stated, increasing the representation of Latinas/os is not enough to influence success.

Implications and Conclusions

This study has implications for higher education policy makers and Latina/o student advocacy groups who need to understand the factors that predict graduation for Latina/o students, especially within HSIs and emerging HSIs. Policy makers have the ability to reallocate capacity building funds to HSIs and to determine policies that will support the advancement of these institutions. Currently, HSIs can apply for capacity building grants offered by the Department of Education, but these monies need to be continually reallocated and increased to have a long-term impact on HSIs’ ability to graduate Latina/o students. Policy makers need to consider the important influence that institutional resources have on graduating Latina/o students. This study can help them to better understand the significance of institutional resources in predicting graduation rates for Latina/o students and help them make more informed decisions about funding and allocations for HSIs.

Higher education researchers should also be concerned with the growth of HSIs and emerging HSIs and the lack of empirical research on these institutions. If we do not study and understand HSIs and emerging HSIs, we will be ignoring the growth of a population of institutions that serves a diverse group of students. Future research must continue to challenge the importance of selectivity in graduating Latina/o students while continuing to explore the role that Latina/o representation on college campuses may have on graduating students. Administrators of large national data sets must also be intentional about sampling students who attend HSIs and emerging HSIs to collect enough data to be used in advanced quantitative methods. This study, for example, would be further enhanced by the use of multilevel modeling that can account for student-level predictors of graduation such as socioeconomic status, precollege experiences, and college level involvement.

As the number of HSIs and emerging HSIs continue to grow, they have the potential to increase the success and social mobility of Latina/o students while also enhancing the diversity and effectiveness of the system of higher education as a whole. A recent report by the White House Initiative on Educational Excellence for Hispanics (2011) contends that HSIs are vital to ensuring the educational success of Latina/o college students. In thinking about the ability of HSIs to increase the success of students within postsecondary education, we must be mindful of the inherent nature of status and hierarchy within the system of higher education, often defined by selectivity and resources, and continue to challenge the status quo. Institutions with lower selectivity and fewer resources can in fact become important players within a system of higher education faced with serving the most diverse student population to date, but we need to continue to study and support the development of these institutions.

Footnotes

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

Author Biography

Gina A. Garcia is a doctoral candidate in Higher Education and Organizational Change at the University of California, Los Angeles (UCLA). Her research interests are in equity and diversity, organizational culture and identity, Hispanic Serving Institutions (HSIs), and Latina/o college student experiences and success.

References

Allison

P. D.

(2002). Missing data. Thousand Oaks, CA: Sage.

Alon

S. A.

Tienda

(2005). Assessing the “mismatch” hypothesis: Differences in college graduation rates by institutional selectivity. Sociology of Education, 78, 294-315.

Bentler

P. M.

(1988). Causal modeling via structural equation systems. In Nesselroade

J. R.

Cattell

R. B.

(Eds.), Handbook of multivariate experimental psychology (2nd ed., pp. 317-335). New York, NY: Plenum.

Bentler

P. M.

(2006). EQS 6 structural equations program manual. Encino, CA: Multivariate Software, Inc.

Bowen

W. G.

Bok

(1998). The shape of the river: Long-term consequences of considering race in college and university admissions. Princeton, NJ: Princeton University Press.

Byrne

B. M.

(2008). Structural Equation Modeling with EQS: Basic concepts, applications, and programming (2nd ed.). New York, NY: Routledge.

Contreras

F. E.

Malcom

L. E.

Bensimon

E. M.

(2008). Hispanic-serving institutions: Closeted identity and the production of equitable outcomes for Latino/a students. In Gasman

Baez

Turner

C. S. V.

(Eds.), Understanding minority-serving institutions (pp. 71-90). Albany, NY: State University of New York.

Cragg

K. M.

(2009). Influencing the probability for graduation at four-year institutions: A multi-model analysis. Research in Higher Education, 50, 394-413.

de los Santos

A. G. J.

Cuamea

K. M.

(2010). Challenges facing Hispanic-serving institutions in the first decade of the 21st century. Journal of Latinos and Education, 9, 90-107.

10.

Dempster

A. P.

Laird

N. M.

Rubin

D. B.

(1977). Maximum likelihood from incomplete data via the EM algorithm. Journal of the Royal Statistical Society, 39(1), 1-38.

11.

Gandara

Lopez

(1998). Latino students and college entrance exams: How much do they really count? Hispanic Journal of Behavioral Sciences, 20(1), 17-38.

12.

Gansemer-Topf

A. M.

Schuh

J. H.

(2006). Institutional selectivity and institutional expenditures: Examining organizational factors that contribute to retention and graduation. Research in Higher Education, 47, 613-642.

13.

Hagedorn

L. S.

Chi

Cepeda

R. M.

McLain

(2007). An investigation of critical mass: The role of Latino representation in the success of urban community college students. Research in Higher Education, 48(1), 73-91.

14.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6(1), 1-55.

15.

Hurtado

Kamimura

(2003). Latina/o retention in four-year institutions. In Castellanos

Jones

(Eds.), The majority in the minority (pp.138-150). Sterling, VA: Stylus.

16.

Hurtado

Milem

J. F.

Clayton-Pedersen

A. R.

Allen

W. R.

(1998). Enhancing campus climates for racial/ethnic diversity: Educational policy and practice. Review of Higher Education, 21, 279-302.

17.

Hurtado

Milem

J. F.

Clayton-Pedersen

A. R.

Allen

W. R.

(1999). Enacting diverse learning environments: Improving the campus climate for racial/ethnic diversity in higher education (Vol. 26). San Francisco, CA: Jossey Bass.

18.

Kane

(1998). Racial and ethnic preferences in college admissions. In Jencks

Phillips

(Eds.), The Black-White test score gap (pp. 431-456). Washington, DC: Brookings Institution Press.

19.

Laden

B. V.

(2004). Hispanic-serving institutions: What are they? Where are they? Community College Journal of Research and Practice, 28, 181-198.

20.

MacCallum

R. C.

Browne

M. W.

Sugawara

H. M.

(1996). Power analysis and determination of sample size for covariance structure modeling. Psychological Methods, 1(2), 130-149.

21.

Malcom

L. E.

(2010). Hispanic-serving or Hispanic-enrolling? Assessing the institutional performance of public 4-year HSIs and emerging HSIs. Paper presented at the American Educational Research Association Annual Meeting. Denver, CO.

22.

McLachlan

G. J.

Krishnan

(1997). The EM algorithm and extensions. New York, NY: Wiley.

23.

Melguizo

(2008). Quality matters: Assessing the impact of attending more selective institutions on college completion rates of minorities. Research in Higher Education, 49, 214-236.

24.

Mercer

C. J.

Stedman

J. B.

(2008). Minority-serving institutions: Selected institutional and student characteristics. In Gasman

Baez

Turner

C. S. V.

(Eds.), Understanding minority-serving institutions (pp. 28-42). Albany: State University of New York.

25.

Mulaik

S. A.

(2009). Linear causal modeling with structural equations. Boca Raton, FL: Chapman & Hall/CRC.

26.

Ryan

J. F.

(2004). The relationship between institutional expenditures and degree attainment at baccalaureate colleges. Research in Higher Education, 45(2), 97-114.

27.

Santiago

D. A.

Andrade

S. J.

(2010). Emerging Hispanic-serving institutions: Serving Latino students. Washington, DC: Excelencia in Education.

28.

Shin

J. C.

(2010). Impacts of performance-based accountability on institutional performance in the U.S. Journal of Higher Education, 60(1), 47-68.

29.

Torres

(2006). A mixed method study testing data-model fit of a retention model for Latino/a students at urban universities. Journal of College Student Development, 47, 299-318.

30.

White House Initiative on Educational Excellence for Hispanics. (2011). Winning the future: Improving education for the Latino community. Retrieved from http://www.whitehouse.gov/sites/default/files/rss_viewer/WinningTheFutureImprovingLatinoEducation.pdf

31.

Zhang

(2009). Does state funding affect graduation rates at public four-year colleges and universities? Educational Policy, 23, 714-731.